Distributed privacy preserving matchmaking protocol based on dynamic interest level in mobile social networks

Making friends based on the similarity of the profile of users is becoming common in many applications of mobile social networking. A major challenge for profile matching is to ensure the privacy of personal profiles which often contain highly sensitive information related to interests, political tendency, health conditions, and so on. The existing privacy preserving matchmaking protocols select the best match among people in the proximity based on the number of common attributes. This lucks detail information to find the best match such as in which attributes of the profile the user is currently interested most. On the other hand schemes that select best match based on the interest level of the user for the attributes assumes the attributes of the users are public which cannot be applied in applications that involve sensitive information about the user. Thus in this paper we propose privacy preserving match making protocol based on the user attributes in the profile and the dynamic interest level of the user for each attribute by applying homomorphic encryption techniques. The best match selected by the initiator have greater than a minimum number of attributes the initiator interested most currently in addition to relatively having a large number of common attributes. Non common attributes of the two matches remain private. Users other than the best matches can at most get the number of common attributes and whether the sum of common attributes interest level is above the threshold. The analysis shows that our protocol is efficient with the ability to resist the semi-honest and malicious adversary. And the simulation result shows that our proposed solution out performs the existing solution in energy overhead and execution time.